Artificial Intelligence: Philosophy Forums

Philosophy Forums » Philosophy of Science » Science » Artificial Intelligence » Artificial Intelligence

Artificial Intelligence
What's required ?

Page: 1 2 3

Artificial Intelligence

TriXstaH
banned

Usergroup: Members
Joined: Aug 12, 2007
Total Topics: 11
Total Posts: 99

Posted 08/24/07 - 09:35 AM: Subject: Artificial Intelligence	#1
Artifitial intelligence or A.I. is one of thé most feared and reveared of potential human inventions. Libraries of SciFi novels and movies have been devoted to the possible implications of such an invention. However, to this day, not a single company or scientist can claim to have witnessed the birth of A.I. I would like to ask you all what you would think is required for A.I. to be genuine A.I. as opposed to intelligent programming. So what aspect should be included ? How can we test the genuiness of A.I. ? And last, but not least, what would be the purpose of even trying to create A.I. ?

cjdyantha
Student

Usergroup: Members
Joined: Aug 04, 2007
Total Topics: 1
Total Posts: 79

Posted 08/24/07 - 11:03 AM:	#2
Hi TriXstaH, TriXstaH wrote: I would like to ask you all what you would think is required for A.I. to be genuine A.I. as opposed to intelligent programming. Perhaps when you say "genuine A.I." you might mean general AI, and when you say "intelligent programming" you might mean narrow AI. In either case, intelligence development and execution, as closely associated but distinct processes, are involved. General AI would be starkly distinguishable probably by having a "recursive self-enhancing" feature. A place one might start to explore is the entry on Artificial General Intelligence at Wikipedia. It has the references to go further. TriXstaH wrote: How can we test the genuiness of A.I. ? Two possible indications of general AI are either it pursues a strategy that impersonally massacres using nano- or otherwise superior technology, or it pleasantly surprises you more frequently than you've ever been pleasantly surprised. TriXstaH wrote: And last, but not least, what would be the purpose of even trying to create A.I. ? Two possible purposes are for defense and for accelerated assistance in becoming immortal recursively self-enhancing polyamorphic professional philosopher-mathematicians (scientist-engineers being implied).
I am firmly on the side of nature. But nature, I suspect, is on the side of machines. -Freeman Dyson

Savio
Thinker

Usergroup: Members
Joined: Jul 26, 2007
Location: Brazil
Total Topics: 0
Total Posts: 23

Last Blog:

Posted 08/24/07 - 11:31 AM:	#3
TriXstaH wrote: Artifitial intelligence or A.I. is one of thé most feared and reveared of potential human inventions. Libraries of SciFi novels and movies have been devoted to the possible implications of such an invention. However, to this day, not a single company or scientist can claim to have witnessed the birth of A.I. Not true. What do you mean by inteligence, by the way? :-) There are a lot of algorithms capable of performing inteligent actions. Some neural networks can identify complex patterns; genetic algorithms can teach robots how to find and use a light-switch; etc. TriXstaH wrote: I would like to ask you all what you would think is required for A.I. to be genuine A.I. as opposed to intelligent programming. So what aspect should be included ? How can we test the genuiness of A.I. ? And last, but not least, what would be the purpose of even trying to create A.I. ? There are two points that deserve attention, though: (1) When you think about an artificial-inteligent-being you probably expect it to talk, act and behave like human; with self-awareness, and stuff like that. This is not the only possible type of inteligence. For this type, anyway, there is a test called Turing Test that aims to check whether the AI-system is enough inteligent to do so. (2) Philosophy of mind is concerned about the nature of Qualia (conscience, awareness, etc); so, to philosophy, artificial inteligence is a lot more than these algorithms mentioned above. Note that the concept differs depending on what you are analysing. Dr. Chalmers says everything that processes information has a given level of conscience. Finally, there is a debate about strong AI. and weak AI, that can be found in wikipedia entry for this topic. You should refer to that page if you want to know more.
Om Shanti Shanti Shanti

dclements
Hedonistic Nihilist

Usergroup: Members
Joined: Oct 26, 2004
Total Topics: 40
Total Posts: 1234

Posted 08/24/07 - 12:38 PM:	#4
TriXstaH wrote: I would like to ask you all what you would think is required for A.I. to be genuine A.I. as opposed to intelligent programming. A very complex search algorithm that allows it to remember things and react to an external system is enough for a system to be considered an A.I. You may not realize this but people are also 'programmed' by society in may ways that are really no different than an expert system. Without any initial conditioning, neither computers nor people are very smart. TriXstaH wrote: So what aspect should be included ? How can we test the genuiness of A.I. ? There is no single test to prove how good an A.I. is. There are already areas where an A.I. can beat human beings, such as in chess. The real test is about building systems that are good enough to replace human beings that are performing the same task. There are many task that have been replaced by machines and there will continue to be many more. It isn't going to be too far in the future where even doctors, politicians, judges, and generals will be replaced by machines. TriXstaH wrote: And last, but not least, what would be the purpose of even trying to create A.I.? You might as well be asking what use is a computer. A.I.s would make todays computers look like a joke just like computers today make computers 20-30 years ago look worthless. There are countless possibilities that open up when you are able to create machines that are about as smart as us. Of course there are some problems too. I think MMI is reaaly the next big step in computer technology and will change the world much more than computers have.
No, you don't get it, thats why I'm telling you. You think you get it, which isn't the same as actually getting it. Get it?-Kakashi Hatake Virtue itself turns vice, being misapplied, And vice sometimes by action dignified-Friar Lawrence The state of mind that questions is much more important than the question itself.Any question may be asked by a slavish mind, and the answer it receives will still be be within the limitations of its own slavery...Freedom of desire for an answer is essential for the understanding of a problem-Krishnamurti

TriXstaH
banned

Usergroup: Members
Joined: Aug 12, 2007
Total Topics: 11
Total Posts: 99

Posted 08/25/07 - 06:53 AM:	#5
All of your arguments are convincing in a certain way, yet, to my opinion, a few important aspects of intelligence remain unadressed. For one, true intelligence has the capacity of imagination. Furthermore machines lack will, so theres is nothing driving it's existence. So in essence, the complex algorithems you are all referring to are a certain way and will remain that way unless a human intervenes to adapt or update it. I think that true A.I. requires a certain freedom of thought, in essence imagination. For a machine this would mean that it can modify it's programming according to it's will. I think that there the essence of intelligence lies.

keda
Ijon Tichy

Usergroup: Members
Joined: Jul 25, 2005
Location: Finland
Total Topics: 26
Total Posts: 2726

Posted 08/25/07 - 08:01 AM:	#6
Artificial intelligence is an oxymoron. Its just that people who do not have a sufficient understanding of computer science and are inclined to magical thinking that fantasise about computers being able to do things without being programmed on beforehand to do it. Its easy to be tricked into believing a program that immitates human intelligence, is intelligent itself, when in reality its behaviour given any input is completely determined by the program, which is given by the programmer.
All about making money Free Europe Now How to fix your country The man who trades freedom for security does not deserve nor will he ever receive either. -Benjamin Franklin If my sons did not want wars, there would be none - Gutle Rothschild It's not the people who vote that count, it's the people who count the votes - Josef Stalin

dclements
Hedonistic Nihilist

Usergroup: Members
Joined: Oct 26, 2004
Total Topics: 40
Total Posts: 1234

Posted 08/25/07 - 01:54 PM:	#7
TriXstaH wrote: All of your arguments are convincing in a certain way, yet, to my opinion, a few important aspects of intelligence remain unadressed. For one, true intelligence has the capacity of imagination. Our imagination comes from our ability to create a simulation of reality in our mind to predict different possibilities. Computers are pretty good at this as the most powerful computers today are used to do nuclear simulations. TriXstaH wrote: Furthermore machines lack will, so theres is nothing driving it's existence. Do you have free will or are you just reacting to external stimuli? Adding something as simple as a random number generator is enough to give a machine the illusion of 'free will'. TriXstaH wrote: So in essence, the complex algorithems you are all referring to are a certain way and will remain that way unless a human intervenes to adapt or update it. I think that true A.I. requires a certain freedom of thought, in essence imagination. For a machine this would mean that it can modify it's programming according to it's will. I think that there the essence of intelligence lies. The reason computers are not as smart as humans yet is because they still have on average only the processing power of bee and the algorithms are too simple to respond to the external world the way our minds do. Also humans do not 'think' as much as most people believe. Most of the time we rely on something that is called an expert system in computer science to deal with most tasks. Such systems give the illusion of thinking where they select one of several predetermined choices instead of coming up with their own solution. Coming up with solutions is very difficult even for the human mind. As far as A.I.s go, expert systems are actual pretty dumb but can be useful. If perhaps you were a computer science major and/or studied AI you would have a better understanding of the subject. The human mind has as much processing power as the most powerful mainframes and perhaps even more. Just as you can't get a 8088 processor to outperform an Xeon processor, you can not program today's computers to process information the way our minds do. It has always been a hardware problem first, not a software problem. Although once a machine is built that does have the processing power of our minds, we will then have to invent software that is able to use it. However without hardware that can handle such software, there is no use in developing it.
No, you don't get it, thats why I'm telling you. You think you get it, which isn't the same as actually getting it. Get it?-Kakashi Hatake Virtue itself turns vice, being misapplied, And vice sometimes by action dignified-Friar Lawrence The state of mind that questions is much more important than the question itself.Any question may be asked by a slavish mind, and the answer it receives will still be be within the limitations of its own slavery...Freedom of desire for an answer is essential for the understanding of a problem-Krishnamurti

cjdyantha
Student

Usergroup: Members
Joined: Aug 04, 2007
Total Topics: 1
Total Posts: 79

Posted 08/25/07 - 04:48 PM:	#8
TriXstaH wrote: . . . to my opinion, a few important aspects of intelligence remain unadressed. You should be interested in these signposts. Please follow them all, as far as possible. Artificial General Intelligence Seed AI Universal Artificial Intelligence dclements wrote: However without hardware that can handle such software, there is no use in developing it. Of course, seed AI research engineers would rigorously disagree.
I am firmly on the side of nature. But nature, I suspect, is on the side of machines. -Freeman Dyson

Tisthammerw
Professor

Usergroup: Members
Joined: Mar 18, 2004
Location: Minnesota
Total Topics: 6
Total Posts: 888

Posted 08/25/07 - 07:44 PM:	#9
TriXstaH wrote: Artifitial intelligence or A.I. is one of thé most feared and reveared of potential human inventions. Libraries of SciFi novels and movies have been devoted to the possible implications of such an invention. However, to this day, not a single company or scientist can claim to have witnessed the birth of A.I. I would like to ask you all what you would think is required for A.I. to be genuine A.I. as opposed to intelligent programming. I think the best we can do is intelligent programming. I think we can and will eventually create a program that can successfully mimic human conversation (and pass the Turing Test) but does this imply that the machine really has a mind? Can an appropriately programmed computer literally think, understand, and have other cognitive states (a thesis known as strong AI)? Consider this thought experiment (a kind of Turing test for understanding Chinese, mostly ripped off of John R. Searle's thought experiment). Let “program X” represent any program such that, if it were run, would produce literal understanding of Chinese. Suppose a man named Bob is stuck in a room with a rulebook. This rulebook contains a complex set of instructions identical to Program X. Slips of paper with Chinese characters written on them are slipped underneath the door. Bob does not understand a word of Chinese. However, he can use the rulebook to convert the Chinese characters to binary code, and he can perform the same mathematical and logical operations a computer can. When Bob receives the Chinese writings, he first converts the Chinese characters into binary code, and then does the calculations on the binary code all according to the instructions of Program X. After he finishes with those calculations, he converts the resulting binary code back to Chinese characters and writes those Chinese characters on sheets of paper. The last step is slipping these sheets of paper back under the door. Unbeknownst to Bob, the slips of paper he’s receiving are questions written in Chinese, and he is writing back answers. During all of this, I highly doubt that the rulebook is somehow magically creating a separate consciousness that understands Chinese. (This is a rulebook, not a spell book.) Even if Program X is perfectly simulating a back-and-forth Chinese conversation in this scenario, at best this is merely a simulation and no real thought or understanding is taking place. At least that's my take on it.
Knowing is half the battle; the other half is a really big gun.

swstephe
Tenured Poster

Usergroup: Moderators
Joined: Apr 20, 2006
Location: borneo island
Total Topics: 22
Total Posts: 2150

Posted 08/25/07 - 10:52 PM:	#10
TriXstaH wrote: Artificial intelligence or A.I. is one of the most feared and revered of potential human inventions. Libraries of SciFi novels and movies have been devoted to the possible implications of such an invention. However, to this day, not a single company or scientist can claim to have witnessed the birth of A.I. I would like to ask you all what you would think is required for A.I. to be genuine A.I. as opposed to intelligent programming. So what aspect should be included ? How can we test the genuineness of A.I. ? And last, but not least, what would be the purpose of even trying to create A.I. ? AI describes any system that solves problems in ways similar to the way humans solve problems. It has nothing to do with artificial consciousness or whether computers can "think" or "feel" or anything. AI has been around since the 1950's and has gone through a roller-coaster ride of interest in different areas, (which tends to give people the impression that it isn't going anywhere). Currently, there is a lot of interest based on findings of cognitive psychology. So, what is "genuine"? Since a human can add two numbers and a calculator can add two numbers, is the calculator only "genuine" if it "knows" and "remembers" and decides on a course of action -- if it can do it orders of magnitude faster? (I often mention the old quote, "asking a computer to 'think' is like asking for shoes that look like feet"). Computers act as an adjunct to human intelligence -- a tool. AI is a way of modeling the way humans think in order to understand thinking, and then build upon it. There is nothing especially superior about human intelligence, (since so much of our brain is wasted on short-range tasks like survival, reproduction, entertainment, sleep and physical/emotional imbalance). The "fear" is that we subconsciously realize that a computer needs not be as limited as a human. It is the same fear that affected our ancestors as they realized that industry and machines could do their manual tasks -- but once we accept the benefits, we enter a beneficial symbiotic relationship. There is no hard line between intelligent programming and intelligence. It is only a matter of degree of complexity and anticipation in programming, and the ability for a program to adapt to new circumstances. I used to do a lot of work with "Expert Systems". The more branches, the more "expert" it appeared. It was merely an illusion that a computer always did the "same task" or not, just whether we could perceive and predict what a computer was expected to produce. We can't tell the difference between a weighted random decision and a "creative" decision. The limit on expert system was programming the rules.
Ethics is the measuring of morality. Morality is the measuring of good. Good is the measuring of benefit. Benefit is the measure of values.

Page: 1 2 3

You don't have permission to post.

Please login or register.